Message-ID: <1f74adab-5ed8-450a-853a-f0b4bfe0e751@gmail.com>
Date: Sun, 9 Jul 2023 09:15:34 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:102.0)
 Gecko/20100101 Thunderbird/102.12.0
Subject: Re: Autogenerate some wait events code and documentation
To: Michael Paquier <michael@paquier.xyz>, Andres Freund <andres@anarazel.de>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
References: <69588eaf-02d7-95a8-0da4-3e120de16a59@gmail.com>
 <ZGWrTn//v+WAqMpN@paquier.xyz> <ZKJxZrnFTxoCx8a5@paquier.xyz>
 <ZKJ0tzXSH1LyqI9V@paquier.xyz>
 <0b3cdac7-f078-61a6-76e0-99db7aae78ce@gmail.com>
 <ZKTN//NfDeP1egLW@paquier.xyz>
 <20230705215939.ulnfbr4zavb2x7ri@awork3.anarazel.de>
 <ZKYMfGRK1PIlGiBB@paquier.xyz>
 <20230707011943.qxbj3vj2iwsfgqvq@awork3.anarazel.de>
 <ZKeZVPEBkBxYs1FY@paquier.xyz> <ZKo4b1bDI5+FoTex@paquier.xyz>
Content-Language: en-US
From: "Drouvot, Bertrand" <bertranddrouvot.pg@gmail.com>
In-Reply-To: <ZKo4b1bDI5+FoTex@paquier.xyz>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Archived-At: <https://www.postgresql.org/message-id/1f74adab-5ed8-450a-853a-f0b4bfe0e751%40gmail.com>
Precedence: bulk

Hi,

On 7/9/23 6:32 AM, Michael Paquier wrote:
> On Fri, Jul 07, 2023 at 01:49:24PM +0900, Michael Paquier wrote:
>> Hmm.  If we go down this road I would make the choice of simplicity
>> and remove entirely a column, then, generating the snakecase from the
>> camelcase or vice-versa (say like a $string =~ s/([a-z]+)/$1_/g;),
>> even if it means having slightly incompatible strings showing to the
>> users. And I'd rather minimize the number of exceptions we need to
>> handle in this automation (aka no exception rules for some keywords
>> like "SSL" or "WAL", etc.).
> 
> After pondering more about that, the attached patch set does exactly
> that.

Thanks!

>  Patch 0001 includes an update of the wait event names so as
> these are more consistent with the enum elements generated.  With this
> change, users can apply lower() or upper() across monitoring queries
> and still get the same results as before.  An exception was the
> message queue events, which the enums used "MQ" but the event names
> used "MessageQueue", but this concerns only four lines of code in the
> backend.  The newly-generated enum elements match with the existing
> ones, except for MQ.

> 
> Patch 0002 introduces a set of simplifications for the format of
> wait_event_names.txt:
> - Removal of the first column for the enums.
> - Removal of the quotes for the event name.  We have a single keyword
> for these, so that's kind of annoying to cope with that for new
> entries.
> - Build of the enum elements using the event names, by applying a
> rebuild as simple as that:
> +   $waiteventenumname =~ s/([a-z])([A-Z])/$1_$2/g;
> +   $waiteventenumname = uc($waiteventenumname);
> 
> Thoughts?

That's great and it does simplify the wait_event_names.txt format (and the
impact on "MQ" does not seem like a big deal).

I also noticed that you now provide the culprit line in case of parsing
failure (thanks for that).

  #
-#   "C symbol in enums" "format in the system views" "description in the docs"
+#   "format in the system views" "description in the docs"

Should we add a note here about the impact of the "format in the system views" on
the auto generated enum? (aka how it is generated based on its format)?

Regards,

-- 
Bertrand Drouvot
PostgreSQL Contributors Team
RDS Open Source Databases
Amazon Web Services: https://aws.amazon.com